See the post named "Searching for binary values" or if you need I can post you
my solution.
> Date: Thu, 9 Sep 2010 16:56:43 +0200
> Subject: Re: About jcr:contains for binary content and properties
> From: [email protected]
> To: [email protected]
>
> On Thu, Sep 9, 2010 at 2:43 PM, Paco Avila <[email protected]> wrote:
> > In the Jackrabbit FAQ I see this question:
> >
> > Why doesn't //*[jcr:contains(@jcr:data, 'foo')] return matches for
> > binary content? Extracted text from binary content is only indexed on
> > the parent node of the @jcr:data property. Use jcr:contains() on the
> > nt:resource node.
> >
> > My problem in that this query:
> >
> > //element(*, nt:file)[jcr:contains(jcr:content, 'foo')]
> >
> > will also match jcr:content nodes with the 'foo' text in their
> > properties. For example, if I want to find document with the word
> > "pdf" inside, will also match documents which jcr:mimeType is
> > 'application/pdf'.
> >
> > How can I search only by the binary indexed data?
>
> You would need to hook into the indexing then. I answered a similar
> question somewhere around 2 weeks ago. You should be able to find this
> back in the archive.
>
> Regards Ard
>
> > --
> > OpenKM
> > http://www.openkm.com
> > http://www.guia-ubuntu.org
> >
